Introduction

The ascomycete genus Chrysoporthe includes several important pathogens of Eucalyptus species. Eucalyptus canker caused by Chrysoporthe cubensis is an important disease of commercial Eucalyptus plantations in Brazil, despite several years of improving the crop for resistance to this disease (Ferreira and Milani 2004; Gryzenhout et al. 2009). Canker resistant clones have been selected by means of controlled crosses between species of Eucalyptus (Assis and Mafia 2007; Alfenas et al. 2009). Chrysoporthe cubensis occurs in the Neotropics, in which conditions of high temperature and relative air humidity prevail (Sharma et al. 1985; Boerboom and Maas 1970; Hodges et al. 1976, 1979). This pathogen also causes canker in several other species in the families Myrtaceae and Melastomataceae. Previously, C. cubensis was reported in Brazil only from T. granulosa and Marlierela edulis (Boerboom and Maas 1970; Hodges et al. 1976, 1979; Sharma et al. 1985; Myburg et al. 2003; Rodas et al. 2005; Seixas et al. 2004; Barreto et al. 2006; Gryzenhout et al. 2006).

The taxonomic positioning of Chrysoporthe cubensis has been problematic since its discovery in 1917, and only in 2004 the species was accommodated within the genus Chrysoporthe (Gryzenhout et al. 2004). The fungus was originally described as Diaporthe cubensis Bruner (1917), but changed to Cryphonectria because of the cultural and morphological characteristics that resembled the latter (Hodges 1980). However, molecular studies based on the LSU and SSU gene sequences clearly showed that C. cubensis was phylogenetically distant from other species within Cryphonectria (Zhang and Blackwell 2001; Castlebury et al. 2002; Gryzenhout et al. 2004).

Based on morphological differences and DNA sequences, new species of Chrysoporthe were described, and the genus currently comprises eight species. In 2010, a new species (C. deuterocubensis) was described to accommodate Asian isolates, which were previously placed in Cryphonectria cubensis (van der Merwe et al. 2010). Two other species, both previously placed in Cryphonectria cubensis, were described as Chrysoporthe doradensis and Chrysoporthe inopina (Gryzenhout et al. 2005; Gryzenhout et al. 2006). The description of both C. doradensis and C. inopina was supported using morphological features which are clearly distinct from C. cubensis (Gryzenhout et al. 2005, 2006). Gryzenhout et al. (2006) found that isolates obtained from Tibouchina spp. in Colombia had only the asexual state, for which they placed in a new genus Chrysoporthella, as Chrysoporthella hodgesiana. Two other species from Zambia pathogenic to Eucalyptus, also known only from their asexual states, were described as Chrysoporthella zambiensis and Chrysoporthella syzygiicola (Chungu et al. 2010).

Reports of new Chrysoporthe species, causing canker in other locations around the world, have increased over recent years (Gryzehout et al. 2009; van der Merwe 2012). These new reports have been made based on a more accurate species delineation partly resulting from the increasing use of molecular tools to characterize isolates. Thus, phylogenetic studies are important to evaluate differences among taxonomic groups and describe novel species from different regions. Only Chrysoporthe cubensis has been reported from Brazil, and little is known about the phylogenetic characterization of Brazilian isolates, despite the importance of canker in Eucalyptus plantations. In addition, there are reports of the fungus on T. granulosa, probably because of its extensive use in urban forestry in Brazil. Although T. heteromalla is used as an ornamental tree in Minas Gerais State (Brazil) and is also naturally found in savanna areas, there are no records of the disease in this species. Species of the genus Tibouchina (Melastomataceae) occur in all Brazilian territory (Souza 1986), mainly in the states of the southeast region (Peralta 2002) and the genus stands out from the economic point of view for its ornamental value. Accordingly, the present study aimed to identify Chrysoporthe species using morphological and DNA sequencing data of β-tubulin and actin regions from Eucalyptus, T. granulosa, and T. heteromalla in Brazil and to determine their pathogenicity to the three hosts.

Materials and methods

Collection of fungal isolates

Stem and bark fragments with typical cankers and fruiting bodies resembling those of C. cubensis were collected from hybrid clones (Eucalyptus grandis x E. urophylla) in Eucalyptus plantations in the Maranhão (MA) and Minas Gerais (MG) states and from T. granulosa and T. heteromalla ornamental trees in Minas Gerais. Isolations were performed by transferring single ascospore and/or conidial masses from pycnidia to Petri dishes containing 39 g/L Potato Dextrose Agar (PDA; HiMedia Laboratories Pvt. Ltd) medium.

Morphological and molecular characterization

Pure cultures prepared on PDA medium were grown for 10 days to obtain sufficient fungal biomass for DNA extraction. Mycelium was scraped from the surfaces of actively growing cultures and macerated in liquid nitrogen, and DNA extractions performed using the Wizard®Genomic DNA Purification Kit (Promega), according to the manufacturer’s instructions.

From the DNA extracted, β-tubulin region was amplified with the primers Bt1a, Bt1b, Bt2a, Bt2b. (Glass and Donaldson 1995) and actin (ACT) region with the primers ACT-512F, ACT-783R (Carbone and Kohn 1999), according to van der Merwe et al. (2010).

All amplification reactions were prepared in a final volume of 25 μl. PCR reactions were performed in a thermocycler (My Cycler™ -BIO-RAD) and conditions were adjusted for each gene as described in Glass and Donaldson (1995) and Carbone and Kohn (1999). The amplified fragments were purified using the Gen Elute PCR Clean-up Kit (Sigma-Aldrich).

Sequencing was performed at the Laboratório de Patologia Florestal, Universidade Federal de Viçosa, and at the Macrogen. The generated electropherograms were edited using SeqAssem Software (Hepperle 2004). Phylogenetic analysis was conducted using a dataset containing sequences of Chrysoporthe from previous studies and obtained from GenBank (Table 1).

Table 1 Information for the isolates of Chrsyoporthe used for phylogenetic analysis in this study
Full size table

Multiple alignments of nucleotide sequences were constructed using the CLUSTALW Program (Thompson et al. 1994) and implemented using MEGA v 6.0 (Koichiro et al. 2013). Sequences were manually edited where needed using SeqAssem (Hepperle 2004). Maximum parsimony (MP) analysis was performed using PAUP* 3.2.1 software (Swofford 2002) (Ronquist et al. 2012) with a heuristic search with 1000 random additions and TBR. A total of 1,978,224 heuristic search replicates, with 100 tres retained. A bootstrap analysis (1000 replicates) was also done on the dataset to determine the confidence levels of the branches using 10 random additions and TBR for each replicate. The gaps in the alignment were treated as missing data. Bayesian inference was used to generate posterior probabilities (PP) for consensus nodes using MRBAYES 3.1 (Huelsenbeck 2001). The analysis was performed using algorithm MCMC (Markov Chain Monte Carlo) (Larget and Simon 1999). Two concurrent analyses of four chains were both run for 107 generations, using random starting trees.

Trees were randomly sampled every 1000 generations, and 25% of them were discarded as burn-in. Trees were viewed and edited in FigTree 1.3.1. (http://tree.bio.ac.uk/software Accession code TreeBase 22716). A combined dataset of ACT, BT1, and BT2 sequences was submitted to a partition homogeneity test (PHT; Farris et al. 1994) to determine whether the datasets could be combined. Amphilogia gyrosa sequences were defined as outgroup taxon according to van der Merwe et al. (2010).

Fungal structures and fruiting bodies from herbarium material were selected based on the results obtained from phylogenetic analysis and used in morphological characterization. For this purpose, histological sections of fungal structures and fruiting bodies were examined according to the method described by Gryzenhout et al. (2004), wherein 50 asci, ascospores, conidia and conidiophores were measured.

Pathogenicity tests

Inoculation of Tibouchina granulosa and T. heteromalla

The pathogenicity of the Chrysoporthe isolates obtained was determined on T. granulosa and T. heteromalla using two isolates, CE6 and CE38 selected according to results of sequence analyses of the β-tubulin and actin regions. Seedlings of T. granulosa (18-months-old) and T. heteromalla (12-months-old) were inoculated and kept under greenhouse conditions. Wounds of 2 cm of length were made with a scalpel by removing the bark to expose the cambium. Subsequently, fungal culture discs (5 mm diameter) were placed in the wound with the mycelium facing the xylem, and then covered with plastic film (Parafilm®) to maintain humidity. This protection was maintained for 30 days after inoculation, as described by Ferreira and Milani (2004). Control treatments consisted of wounded seedlings receiving only a disc of PDA without the fungus. After 75 days, the plants were evaluated by measuring the lesion lenght in the xylem. Koch’s postulate was fulfilled by re-isolating the fungus. Pathogenicity tests were conducted in a completely randomized design, with 10 replicates for T. granulosa plants, three replicates for T. heteromalla, and five replicates for each Eucalyptus clone.

Inoculation of Eucalyptus clones

Pathogenicity tests were carried out in 6-month-old Eucalyptus clonal plants, and inoculation was performed as described above, using 10 different Eucalyptus hybrid clones (Eucalyptus grandis x E. urophylla) (T3, T4, T5, T6, T7, T8, T9, T10, T11 and T12) with the isolates CE6 and CE38. Eucalyptus plants were assessed by measuring the lesion length formed in the xylem at seventy-five days after inoculation. Koch’s postulate was carried out by pathogen re-isolation.

Statistical analysis

The normality of both tests was analyzed using Shapiro-Wilk test (Shapiro and Wilk’s 1965). The means of each treatment were compared and grouped according to the Scott-Knott’s test (p ≤ 0.05). Statistical analyses were performed using SISVAR software (Ferreira 2010).

Results

Collection of fungal isolates

Typical fungal structures were observed on diseased T. granulosa, T. heteromalla, and Eucalyptus plants. Only the asexual state was found on Tibouchina species. In species of Tibouchina, the main symptoms observed were cankers and the death of branches and trunks (Fig. 1). On Eucalyptus sp., we observed thickening of the trunk base, bark cracking, and perithecia and pycnidia of the pathogen on the bark of diseased trees.

Fig. 1
figure 1

Chrysoporthe spp. hosts: dead Tibouchina granulosa tree (a), fruiting bodies under the Eucalyptus bark (b), pycnidium with golden spore mass at the apex (c) and C. doradensis conidia (d). Scale bar: c = 100 μm, d = 2 μm

Full size image

Molecular and morphological characterization

Phylogenetic trees were generated by using the maximum parsimony (MP) and Bayesian inference (BI) methods for the amplified BT1, BT2, and ACT regions, using sequences from 17 isolates from Eucalyptus sp. and Tibouchina spp. Twenty two additional sequences of Chrysoporthe spp. obtained from GenBank were included in the analysis for comparison. The trees generated for two gene regions showed similar topologies in both parsimony and Bayesian methods.

The aligned DNA sequence of the BT1 region had an alignment of 410 characters, 198 constant characters, with 196 parsimony non-informative and 16 parsimony informative, produced 100 trees of 230 steps (Consistency index = 0.991, Retention index = 0.986, and Homoplasy index = 0.008). For BT2, the dataset had an alignment of 389 characters produced 4 trees of 45 steps with 343 constant characters, 36 parsimony non-informative and 11 parsimony informative (Consistency index = 0.977, Retention index = 0.989, and Homoplasy index = 0.022). The alignment of the actin gene region had a total 271 characters, with 235 constant, 29 parsimony-uninformative and 7 parsimony informative characters produced 10 trees of 37 steps (Consistency index = 1.000, Retention index = 1.000, and Homoplasy index = 0.000). Evolution models GTR + G for alignment to BT1, BT2 and ACT were selected and incorporated into the Bayesian analysis. The combined dataset of β-tubulin region (BT1 + BT2) and actin had an alignment of 1067 characters, with 763 constant, 266 parsimony non-informative and 38 parsimony informative characters.

A consensus tree showed that the isolates sequenced in this study were grouped with C. doradensis and C. cubensis. Branch was supported by high bootstrap and high posterior probability values for C. doradensis being of 86% and 0.99, respectively, and C. cubensis with 90% and 0.97, bootstrap and posterior probability values, respectively (Fig. 2).

Fig. 2
figure 2

One of 10 most parsimonious tree of β-tubulin and actin regions (combined dataset of ACT, BT1, and BT2). Isolates in bold were sequenced in this study. The bootstrap and posterior probability values greater than 50% and 0.5, respectively, are shown appropriate branches. Amphilogia gyrosa was defined as outgroup taxon

Full size image

The isolates CE6 and CE38 were selected for morphological characterization based on the results of the phylogenetic analyses. No variation in color and morphology was observed among isolates when grown on PDA medium. Diseased tissue showed abundant black perithecial or pycnidial necks in Eucalyptus and black pycnidial necks on Tibouchina species. Isolate CE6 had the typical morphology of C. cubensis: ascomata perithecial, subglobose, ostioles rostrate, asci unitunicate, clavate (18.5–24.0 × 4.5–6.5 μm); ascospores ellipsoid (4.5–6.0 × 2.0–2.5 μm), one septa and hyaline. Conidiomata were superficial to slightly immersed, fuscous-black bases, pyriform to clavate, with one to four attenuated necks per structure. Conidiomata were occasionally multilocular, with a single locule connected to one or several necks. Conidiophores were hyaline and with globose rectangular basal cells that are 10.0–12.5 × 17.0–21.0 μm), branched irregularly at the base or above the cylindrical cells with total length between, conidiogenous cells cylindrical to flask-shaped with attenuated apices (1.5–2.0 × 2.5–3.0 μm). Conidia were hyaline (2.5–3.0 × 1.5–2.0 μm), oblong, non-septate, exuded as bright luteous spore tendrils or droplets.

Isolate CE38 had the typical morphology of C. doradensis: ascomata perithecial, fuscous-black base and cylindrical perithecial necks; Asci fusoid to elipsoidal, (20.0–24.5 × 4.5–6.5 μm). Ascospores hyaline (5.0–6.5 × 2.0–2.5 μm), one septa, fusoid to oval. Conidiophores hyaline, with globose to rectangular basal cells, branched irregularly at the base, total length of conidiophore 10.0–16 × 19.0–21.5 μm, conidiogenous cells 1.5–2.0 μm. Conidia 3.5–4.5-μm long, 1.5–2.0-μm wide, aseptate, cylindrical, oblong, ovoid and occasionally allantoid (Fig. 1d).

Pathogenicity test on of Tibouchina granulosa and T. heteromalla plants

After inoculation, discoloration progressed vertically in the xylem of inoculated T. granulosa and T. heteromalla plants (Fig. 3a, c). A significant difference was found between isolates of the two Chrysoporthe species tested (P < 0.0001) (Table 2). No lesions were produced on control plants of either host (Fig. 3b, d). Both isolates, CE6 and CE38, were pathogenic when inoculated on T. heteromalla (Table 2). However, the C. cubensis isolate showed longer lesion length on T. granulosa than C. doradensis.

Fig. 3
figure 3

Lesions on xylem of Tibouchina granulosa, T. heteromalla and Eucalyptus plants 75 days after inoculation with Chrysoporthe doradensis. a, c, e lesion length on T. granulosa, T. heteromalla and Eucalyptus, respectively. b, d, f control of T. granulosa, T. heteromalla and Eucalyptus, respectively

Full size image
Table 2 Xylem lesion length (cm) in Tibouchina spp. and Eucalyptus clones in response to inoculation with isolates of Chrysoporthe cubensis and C. doradensis
Full size table

Pathogenicity test on Eucalyptus plants

The inoculation of Eucalyptus plants showed different responses depending on the pathogen species or Eucalyptus clone applied, as well as the interaction between these factors (p < 0.0001) (Table 2). No lesions were produced on control plants of any of the clones inoculated (Fig. 3f). Clones T9 and T11 were highly susceptible to both Chrysoporthe species (Table 2). Clones T4 and T5 showed high resistance to both species. Chrysoporthe cubensis isolate CE6 caused longer lesions on the majority of the tested clones.

Discussion

Canker disease of Eucalyptus sp. caused by C. cubensis has been recognized as a serious disease in Brazil since the 1970’s. The disease has been controlled primarily through the use of resistant species and clones. However, it is still considered to be a serious problem in the North and Northeast of Brazil due to expansion of plantations and lack of screening for resistant clones for these areas. The morphological and molecular analyses presented in this study demonstrated for the first time the occurrence of another species, C. doradensis, in Brazil on Eucalyptus and Tibouchina spp.

Chrysoporthe doradensis was described by Gryzenhout et al. (2004) from Ecuador based on phylogenetic studies using ITS and two regions of the β-tubulin genes. The origin of this species is unknown, although studies are indicating it to be native to Ecuador (Gryzenhout et al. 2005), in part because there are no reports of its occurrence in other regions of South America or elsewhere except for the new report of the fungus from Brazil. However, it is still unclear where the center of origin lies. A broader collection of isolates is needed to provide more detailed information to support evolutionary inferences.

In the current study, the fruiting bodies found on both hosts were similar to those reported for C. doradensis (Gryzenhout et al. 2005), except for the yellowish-brown color of the spore mass, which was similar to that reported for other Chrysoporthe species. Nevertheless, conidial morphology seems to be the most important characteristic to distinguish C. doradensis according to Gryzenhout et al. (2005). These authors observed that conidial shape and spore mass color were the most pronounced morphological characteristics, supporting phylogenetic distinction of Ecuador isolates from other species. Such morphological characteristics were observed in this study, with conidia occasionally allantoid (3.5–4.5-μm long, 1.5–2.0-μm wide) and exuded as bright luteous spore tendrils or droplets.

Host-specialization was not observed among isolates of C. doradensis and C. cubensis collected from Eucalyptus sp. and T. granulosa. The presence of C. doradensis in Brazil demonstrates that species is not restricted to only one country, as it was thought to occur only in Ecuador, nor to a single host. This report is highly relevant in the management of the disease and in breeding programs. Both species, C. cubensis and C. doradensis were recovered from trees with high proximity to each other, due to this, great care must be taken in the identification of these fungi, since studies have demonstrated the high genetic diversity present in populations of C. cubensis in Brazil (van Zyl et al. 1998). In addition, the occurrence of these two species suggests that, due to the great territorial extension that the country presents, probably other species of Chrysoporthe may occur in native plants in Brazil, being pathogenic to Eucalytus and presenting a risk to the commercial plantations of the country. However, more studies with more extensive collections would be important to more accurately identify the distribution of Chrysoporthe in Brazil. In addition, this study proved that the diversity of hosts of Chrysoporthe species, implying serious disease management issues, such as the risk of commercial eucalyptus plantations being close to areas of native Chrysoporthe hosts. Based on the results of this study, further phylogenetic analyses, including more isolates, will be important for clarifying the origin and more accurately describing the geographic distribution of the species, especially in South America.

The pathogenicity tests demonstrated that isolates of C. cubensis and C. doradensis are able to cause canker and death in T. granulosa and T. heteromalla plants. Previous studies demonstrated that Eucalyptus sp. is relatively more susceptible than T. granulosa (Leppik 1970; Rodas et al. 2005; Gryzenhout et al. 2005). It is believed that pathogens are less virulent on their native hosts than to exotic species (Newhouse 1990; Liu and Milgroom 2007) which corroborate our finding that T. granulosa (which is native to Brazil), had smaller lesion length when inoculated with isolates from both fungal species.

Findings in the present study also corroborate other studies that have evaluated resistance in Eucalyptus sp. based on lesion length in the stem with relative susceptibility (Guimarães et al. 2010). The search for more efficient and accurate methods to determine the resistance level of Eucalyptus clones to canker diseases is quite important for its control in Brazil (Alfenas et al. 2009). The existence of genetic variability in Eucalyptus enables controlling the disease by selecting resistant materials (Assis and Mafia 2007; Alfenas et al. 2009). This strategy might be also applied for T. granulosa plants, considering the inoculation method adopted for use in this study.

In general, isolate CE6 of C. cubensis had the longest lesion length according to pathogenicity tests for both hosts. Isolate CE38 of C. doradensis resulted in shorter lesion length in Eucalyptus clones when compared with the C. cubensis isolate. By contrast, pathogenicity studies conducted by Gryzenhout (2005) showed that C. doradensis was highly aggressive to E. grandis species. However, in the present study, we evaluated clones of the hybrid E. grandis × E. urophylla, which may explain the difference between the studies.

The present study proved that C. doradensis and C. cubensis were pathogenic to Eucalyptus and Tibouchina. In addition, this study reports the occurrence of C. doradensis as the etiologic agent of canker disease in Eucalyptus and T. granulosa for the first time in Brazil. Nevertheless, more isolates of both species must be used in future studies to better understand the level of susceptibility in these plants.